I think this is the “correct” Bitcoin model. Hybrid log and linear periodic power law model:

Quantoshi

I think I’ve found a good bitcoin price model. The trick is to detrend by extracting a power law floor first, then fit a hybrid damped log periodic and an undamped linear periodic function to the residuals.

Quantoshi

Using a conservative bitcoin price model projected out further than is prudent, a return of 22% per year, every year (or average thereof) for _20 years_ is not unreasonable. This is a >50x return.

Quantoshi

Bitcoin is for the living…or at the very least it doesn’t fund its own destruction…it should be noted that making quantum vulnerable coins valid for CRQC harvesting is the equivalent funding self destruction. I predict people will reject a dead bitcoin in favor of a living bitcoin and thus we will sunset public key exposed outputs within the next 20 years.

My Bitcoin bubble model predicts a boring year. 5.9% CAGR for the next year. Trough of 17% down from here by May. Interact here:

Quantoshi

Trying to get AI to work properly with retirement, bitcoin, conventional financial assets, retirement accounts, and taxes is really hard. So many rules AI is having a very hard time figuring out what to sell and how much tax to pay and opportunity cost of selling…I’m doing magic numbers for time horizon of opportunity cost, but in reality it should be a simulation derived parameter. Anyhow, you know something is fundamentally no good if it’s so complicated nobody can comprehend it.

So I’ve been vibe coding up this bitcoin price extrapolator and have been adding new features as I go…well, I added a DCA accumulation extrapolator and then did a Saylor style borrow to accumulate option and then a retirement simulator using the Monte Carlo method over a Markov Chain…and then added different price models and then let users draw their own model over the price data …anyhow, I decided to try to do a multi asset retirement simulator which was challenging to get working…and then I tried to model taxes in the retirement section. And wow did this get complex. This feature will probably be my last and will take the longest to implement and test. At least a month. Taxes are way more complicated than pre computing 1 GB of MCMC results and setting up a 3 layered cache system for serving plots and making tab switching appear seamless… You’ll find it on the Citadel Planner tab at

Quantoshi

when it’s done.

The slope of the line that goes beneath almost every bitcoin price point ever had a slope of almost exactly 5 on a log log graph with proper t=0. Expect bitcoin price 2^5=32 times higher in 17 years or so.

Quantoshi

see user defined model

If bitcoin price were not to drop lower, this would be the fastest trip from ATH to trough in Bitcoin history.

Quantoshi

I think I have a unique and genuine insight: bitcoin is following a power law with exponent 5 and bubbles superimposed on top making the exponent look closer to 6. A lot to explain, but summary analysis at

Quantoshi

Support power law model + bubble modeling or some periodic component are areas worth investigating. This think I’m getting fairly close with my bubble model; interestingly, there seems to be a relationship between area under the curve of the bubbles and power law exponent…high exponent predicts small bubbles and small exponent predicts larger bubbles…but floor and bubble ceiling converging.

Another important factor to consider in power law analysis: when do we start? Setting a proper zero time makes a ~100 fold difference in value when doubling bitcoin lifetime.

Exponential vs. Powerlaw…PlanB’s Stock to Flow model is very useful over a certain range. But because it is exponential, there will come a time when it is not useful. Power law models can stay useful over a much broader range. It’s not that Plan B is “wrong,” it’s that the usefulness of his model has an expiration date. And he knows this.

The father of Alexander the Great commissioned a servant to remind him daily: “Phillip, you will die.” Memento mori.